SWE Intern @ Doppel
Software Engineer at Doppel. EECS at UC Berkeley.
Studying EECS at UC Berkeley on a Regents Scholarship. I care about making ML systems harder to fool, shipping products with real users and real constraints, and security tooling that teams actually run in production. At Doppel, I work on alert routing interpretability and DNS tooling for 150+ enterprises. With Prof. Nidhi Rastogi at AI4Sec Lab, I researched concept-based defenses against adversarial patch attacks. I also founded my high school's first RoboSub AUV team and competed at CyberPatriot nationals.
Selected projects across research, security, and product — with case studies where there's enough depth to go deeper.
Institutional-grade trading card exchange with Supabase Realtime price feeds, Stripe Connect KYC settlement, and an AI intake pipeline using Gemini Vision and PSA's API to onboard graded cards instantly.
Social engineering defense platform used by Robinhood, Coinbase, OpenAI, and more. Shipped alert routing interpretability to 150+ enterprises and rebuilt DNS tooling recovering ~3,500 false positives yearly.
With Prof. Nidhi Rastogi. Concept-based masking for adversarial patch defense — accepted at NeurIPS 2025 Reliable ML workshop.
Built trading card exchange end-to-end: $22k+ volume, 710+ trades, Stripe Connect settlement, Gemini Vision + PSA API intake.